Long-term chronic alcohol abuse is associated with structural brain changes and neuro-cognitive impairment. Few studies have shown a convincing correlation between these phenomena and it appears likely that neurosubstrates other than structural alterations underlie the cognitive changes associated with heavy drinking and recovery. The overall goal of this project is to test the hypothesis that axonal/dendritic and membrane phospholipid ( and possibly perfusion) changes in white matter underlie the reversible structural and neurocognitive changes associated with long-term chronic alcohol abuse and recovery. Subjects: 50 light drinkers (LD) and 100 heavy drinkers. LD will be studied at baseline and 9-12 months later, HD will be studied at entry into alcohol abuse treatment (to capture the full extent of brain damage due to heavy drinking), at 2-4 weeks of abstinence, and at 9-12 months after treatment entry during abstinence or relapse. Measurements: Cognition by neuropsychological testing; brain structures by MRI, axonal/dendritic and neuronal viability by 1H MR spectroscopic imaging (N-acetyl aspartate, a putative neuronal/axonal marker); lipids by choline-containing compounds (Cho) and myo-inositol (ml) and by phosphorus-31 MRS (via membrane phospholipids and their breakdown products and precursors); regional cerebral blood flow will be measured with exploratory spin-tagged perfusion MRI. The specific focus of the study will be on white matter, but cortical and subcortical gray matter, cerebellum, hippocampus, corpus callosum, and brain stem, intracranial volume and volumes of various brain nuclei will also be assessed. We expect that initially low regional NAA and phospholipid measures and initially high Cho and mI measures correlate with specific measures of cognitive impairment and that these outcome measures will recover during abstinence in association with cognitive improvements; relapse will arrest structural, metabolic, and cognitive improvements. The significance of these results is several fold: First, this project will develop non invasive outcome measures which provide objective quantitative measurements of alcohol-induced brain damage. This may be useful in future clinical trials in which drugs or treatments are used to reduce drinking, or to monitor effects of drugs aimed at reducing brain damage, or facilitating recovery. Second, these results may also provide information, which can lead to the development of specific drug treatments, aimed at preventing brain damage at the neuron or membrane or at facilitating recovery.